Rudnizki et al.

Integrating Materials and Manufacturing Innovation 2012, 1:3

http://www.immijournal.com/content/1/1/3

RESEARCH

Open Access

Phase-field modelling of microstructure evolution

during processing of cold-rolled dual phase steels
Jenny Rudnizki, Ulrich Prahl* and Wolfgang Bleck
* Correspondence: ulrich.prahl@iehk.
rwth-aachen.de
Department of Ferrous Metallurgy,
IEHK, RWTH Aachen University,
Intzestr 1, 52056, Aachen, Germany

Abstract
Cold-rolled dual-phase steels, which belong to the advanced high strength steels,
have gained much interest within the automotive industry. The formation of dualphase microstructure, which provides an optimal combination of strength and
formability for automotive applications, occurs during intercritical annealing of coldrolled strip. Variations in the chemical composition as well as in the heat treatment
parameters influence very strongly the microstructure development and therefore
the final mechanical properties of the strip. Thus, the precise control of the
microstructure evolution during full processing route is required for the achievement
of essential mechanical properties. The current work is focused on a through-process
model on a microstructural scale for the production of dual-phase steel from coldrolled strip, which is based on Phase-Field Method and combines the description of
ferrite recrystallisation and all phase transformations occurring during intercritical
annealing. This approach will enable the prediction of final microstructure for varying
composition and processing conditions, and therefore, can be used for the process
development and optimisation.
Keywords: Phase-field modelling, Phase transformation, Dual-phase steels

Background
The mechanical properties of dual-phase steels are provided by a microstructure which
consists of two different phases with different hardness, i.e. hard martensitic islands in
a soft ferrite matrix. Thereby, not only phase fractions but also distribution and
morphology of both phases are responsible for the mechanical properties [1,2]. The
formation of dual-phase microstructure from cold-rolled strip occurs during intercritical annealing which involves different metallurgical phenomena, like recrystallisation
and phase transformations [3,4]. Each of these processes contributes to the establishment of the final microstructure. Thus, the control of the microstructure evolution
during the whole processing of dual-phase steels is therefore the key for the processing
of steels with predetermined mechanical properties. Therefore, several detailed experimental studies of DP microstructure evolution have been performed e.g. [5-7].
Beside of that, the numerical investigation which is an alternative to complex and expensive experiments can be very helpful. Nowadays, computational materials science is
a powerful approach for understanding physical mechanisms and their interactions
during industrial processing allowing the prediction of microstructure as well as of
2012 Rudnizki et al.; licensee Springer. This is an Open Access article distributed under the terms of the Creative Commons
Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any
medium, provided the original work is properly cited.

Rudnizki et al. Integrating Materials and Manufacturing Innovation 2012, 1:3

http://www.immijournal.com/content/1/1/3

Page 2 of 13

mechanical properties [8,9]. It can be applied for the improvement of process parameters and optimisation of chemical composition.
Up to now, different mechanisms along the annealing route of cold-rolled strips in
dual-phase processing, as there are recrystallization in cold formed ferrite, austenite
formation, grain growth in ferrite and austenite, formation of ferrite and martensite,
have been investigated separately. However, the current trend in computational materials science moves towards ICME (Integrative Computational Materials Engineering)
where individual approaches on the different scales are combined in order to describe a
process chain completely [10,11]. Such combination of models describing single process
steps allows the prediction of final product properties from the knowledge of chemical
composition and processing conditions.
At the time, no through-process model for the processing of dual-phase steels, which
covers all metallurgical phenomena on a microstructural scale in one integral approach,
is available. Setting up such through-process model is expected to lead to significant
improvement in the prediction of microstructure development and therefore, to enable
robust production of dual-phase steels with predefined mechanical properties. Hereby,
the Phase-Field approach represents a powerful method for the calculation of microstructure evolution in multi-phase multicomponent systems which can be applied to
the establishment of through-process model. The potential and ability of this approach
for the prediction of several phase transformations have been already show in many
works [12-15].
The main focus of this work is the formulation of a through-process model on a
microstructural scale for the production of dual-phase steels from cold-rolled strip by
means of Phase-Field approach. 2D- and 3D-simulation results of the microstructure
evolution during intercritical annealing will be shown and compared to experimental
results.

Methods
Experimental
Material characterisation

The material used in this study is industrially produced cold-rolled steel sheet of
1 mm thickness and 60% cold rolling reduction. The chemical composition of the
steel is listed in Table 1. As in a typical DP-steel, the main alloying elements are
C, Si, Mn, and Cr.
According to metallographic analysis, the initial microstructure of investigated material consists of two phases: the bright ferrite matrix and dark pearlite grains, Figure 1.
Both types of grains have an elongated shape, due to cold-rolling. The average pearlite
content was determined to be 16% by means of image analysis. In order to determine
the distribution of the alloying elements in the initial structure, a microprobe (ESMA)
analysis was performed. The concentration of C, Si, Mn, and Cr was measured along
the arrow in the zone between two indents that were used to mark the selected area.
Table 1 Chemical composition (in wt.-%) of the investigated steel
C

Si

Mn

Cr

Al

Nb

Ti

0.098

0.24

1.65

0.017

0.007

0.55

0.04

0.004

0.001

0.002

0.008

Rudnizki et al. Integrating Materials and Manufacturing Innovation 2012, 1:3

http://www.immijournal.com/content/1/1/3

Figure 1 Nital-etched micrograph of initial structure with two indents (left) and ESMA-results (right).

According to the microprobe analysis, carbon concentration of pearlite is about 0.60

wt.-%. In parallel, Thermo-CalcW calculations were performed taking into account C,
Si, Mn and Cr. This confirms as well the eutectic concentration of investigated steel at
approximately 0.60 wt.-% of carbon. The C concentration in ferrite is much lower.
However, the exact determination of the carbon content in ferrite by means of microprobe analysis is limited due to the very small concentration. No difference was found
between the concentrations of Si, Mn and Cr in ferrite and pearlite in the initial structure: the content of the substitutional elements in both phases is equal to the total content in the steel. Thus the initial structure contains 16% of pearlite with 0.60 wt.-% C,
0.24 wt.-% Si, 1.65 wt.-% Mn, and 0.55 wt.-% Cr in a ferritic matrix with the same content of substitutional elements. Knowing the carbon content and phase fraction of
pearlite, the carbon concentration in ferrite was calculated to be 0.002 wt.-% using total
carbon content and mass balance.
The stored energy distribution, which is also an important characteristic of material,
was obtained based on EBSD analysis assuming that the misorientations within deformed
grains reflect the distribution of a small angle grain boundary and therefore, the stored
energy. The approach to identify the stored energy from misorientation maps has been
developed by Zaefferer and can be found at [16,17]. The average value of the stored energy
for the investigated steel was evaluated to be 4.5 J/cm3. Additionally, the EBSD analysis is
applied in this work for the determination of ferrite grain size distributions.
Experimental process simulation

In order to investigate the microstructure evolution during intercritical annealing, tests

were performed on Baehr Dilatometer DIL-805A/D. For these experiments, the samples with 1 mm thickness, 4 mm width and 8 mm length were cut parallel to the rolling
direction of the ingot. The dilatometric tests have been carried out with a dilatometer
type DIL-805A/D fabricated by Baehr Thermoanalyse GmbH. Experiments were realised by a low-pressure environment of the order of 10-5 MPa in order to protect the
samples from oxidation. During experiments temperature was recorded by a thermocouple. Figure 2 represents intercritical heat treatment performed in dilatometer which
corresponds to a simplified industrial processing of dual-phase steel. According to this
heat treatment, the samples were heated in two-steps to the intercritical temperature of
800C, held 60 s at 800C and afterward cooled down slowly with 1C/s to 600C.
Along the intercritical heat treatment samples were quenched at various stages as
can be found in Figure 2 using a helium atmosphere. Thus, the microstructure at the

Page 3 of 13

Rudnizki et al. Integrating Materials and Manufacturing Innovation 2012, 1:3

http://www.immijournal.com/content/1/1/3

Page 4 of 13

Temperature, C

1000

800 C - 60 s

800

1 C/s

1 C/s

600
400

10 C/s
200
0
0

100

200

300
Time, s

400

500

600

Figure 2 Intercritical heat treatment processed in dilatometer; at various stages process was
interrupted by quenching in order to freeze the microstructure.

chosen temperatures was frozen and afterwards studied by means of light optical microscopy. The corresponding average values of the ferrite fraction were obtained by an
image analyser of four nital-etched micrographs per sample at magnification 1000x.
Modelling
Multicomponent multiphase-field approach

The Phase-Field approach is developed for the modelling of the phase transformations
in multicomponent systems [18-20]. For simulation work in this paper the commercial
software package MICRESSW is used which allows the calculation of phase fractions
during solid-solid transformations in multicomponent steels as well as the description
of the corresponding microstructure evolution [21].
Multiphase-Field is a computational approach which describes the evolution of multiple Phase-Field parameters i ! x; t in time and space. These fields map the spatial
distribution either of different grains with different stored energy or of phases with different thermodynamic properties. At the interfaces, the Phase-Field variables change
continuously between 0 and 1 over an interface thickness which can be chosen to be
large compared to the atomic interface thickness, but small compared to the microstructure length scale. The time evolution of the phases is calculated by a set of PhaseField equations formulated by minimization of the free energy functional [19]:

_ i



p
 !
!
!
i j Gij c; T
Mij n ij n Kij
j

Kij j r2 i  i r2 j


2 
i  j
2

In Eq. (1), ij is the mobility of the interface as a function of the interface orientation,
given by the normal vector n!. *ij is the effective anisotropic surface energy (surface
stiffness), and Kij is related to the local curvature of the interface. The interface is
driven by the curvature contribution *ij Kjj and by the thermodynamic driving force
Gij . Gij , which is a function of the local composition ! c, couples the Phase-Field
equations to the diffusion equations:

Rudnizki et al. Integrating Materials and Manufacturing Innovation 2012, 1:3

http://www.immijournal.com/content/1/1/3

!
_c r

N
X

!
i D i r !
ci ;

Page 5 of 13

i1

where ! D i is the multicomponent diffusion coefficient matrix for phase . ! D i is calculated online for the given concentration and temperature using the Fortran TQW
interface to Thermo-CalcW and the mobility databases MOB2 [22]. In case of
recrystallization modelling the thermodynamic driving force is replaced by stored
energy.
Input parameters

2D-Modelling The performed 2D-Phase-Field simulation is carried out based on the

real microstructure of cold-rolled steel, Figure 3. Here, the left image represents the
nital-etched micrograph of the cold-rolled sample, where the ferrite is bright and pearlite and grain boundaries are dark. The centre image shows the phase arrangement
reproduced by simulation software. The size of this 2D-simulation domain is about
110 80 m2. Similar to the nital-etched micrograph, in the replicated pattern, ferrite is
white while pearlite and grain boundaries are black. The right image shows the carbon
distribution in black-and-white scale, wherein black corresponds to zero and white to
the maximum carbon concentration 0.7 wt.-%.
There, the important material characteristics, i.e. distribution of alloying elements
and stored energy, are taken from the microprobe and EBSD analyses. In the simulation
C, Si, Mn and Cr were included, which are the main alloying elements in this DP600
grade. Thus, according to the microprobe results, it is taken that the substitutional elements are homogeneously distributed in the structure whereas the carbon content in
pearlite is about 0.60 wt.-% and in ferrite 0.002 wt.-%.
The stored energy of ferritic grains which is the driving force for the recrystallisation
is set between 3 and 7 J/cm3, so that the average value corresponds to the EBSD results.
As pearlite is a much harder phase compared to ferrite, the stored energy of pearlite is
set to zero and thus, recrystallisation occurs only in ferrite and is finalised before austenite forms.
Recrystallisation is followed by the nucleation and growth of new grains. Due to the
high heating rate in industrial annealing lines, recovery has been neglected in this
study. Additionally to the stored energy, which is the driving force for this process, the

Figure 3 Nital-etched micrograph of the cold-rolled sample (left), phase arrangement reproduced
by simulation software (center) and representation of carbon distribution in black-and-white scale
(right).

Rudnizki et al. Integrating Materials and Manufacturing Innovation 2012, 1:3

http://www.immijournal.com/content/1/1/3

main input parameters for the simulation of recrystallisation are nucleation density and
grain boundary mobility. These parameters determine the rate and characteristics of
the whole process.
To approximate the real transformation behaviour, 2D-simulation of recrystallisation
is performed following a local maximum stored energy criterion assuming continuous
nucleation. The reason for this is the fact that in the 2D-simulations each grain emerging on the simulated surface should be assumed to be a newly-generated. Thus, for
the 2D-modelling it is important during the whole process to introduce new grains
which reflect and compensate growth of grains from the underlying level in the considered 2D-section. The total number of nuclei appearing during the simulation is set to
250 based on the experimental grain size distribution after recrystallisation. This approach is similar to that used by [23] combining nucleation scenario assumptions with
mobility adjustment to replicate the experimental data.
The grain boundary mobility is typically given by the Arrhenius equation describing
its temperature dependence. The activation energy for / grain boundary mobility is
fixed to 140 kJ/mol as this is the most widely used value in the literature [24]. The preexponential factor is set to 3.2 10-6 m4/(Js) based on the comparison of the simulated
growth of the single grains by the variation of the pre-exponential factor with the experimental grain sizes. As the simulation focuses on micro scale, all grain boundaries
are assumed to be high angle grain boundaries and will be kinetically treated
identically.
The modelling of phase transformation (austenite formation from ferrite pearlite and
ferrite formation from austenite) will be performed separately and the results will be
combined in one through process simulation. For the simulation of pearlite dissolution
a simplified approach is used. Pearlite in this approach is considered as an effective
pseudo-phase with mixed properties of ferrite and cementite. The thermodynamic driving force for dissolution is calculated from overheating obtained by a linearization of
the phase diagram. The interactions of pearlite with other phases are defined by
assumed slopes which restrict the two phase regions [13]. Based on the metallographic
observations, nucleation of austenite during heating is assumed to take place only
within pearlitic grains. The number of austenite nuclei is set to 100 so that in the large
pearlitic grains several austenite nuclei could form, similar to the metallographic
finding.
The thermodynamic driving force for the modelling of ferrite-to-austenite and
austenite-to-ferrite transformations is obtained by the direct coupling to the thermodynamic database via the Gibbs energy minimisation software Thermo-CalcW. The
modelling of these phase transformations are performed taking into account redistribution of substitutional alloying according to LENP (Local Equilibrium Non Partitioning)
conditions using special model for solute redistribution [12].
The diffusion parameters are taken from the kinetic database MOB2. The mobility
for the ferrite/austenite and pearlte/austenite grain boundaries are defined by an activation energy, of 140 kJ/mol and a pre-exponential factor of 1.5 10-6 m4/(Js). The other
input parameters which have to be mentioned here are interfacial energy the interfacial
energy *ij, of 0.4 Jm-2 for ferrite/austenite interface, 0.9 Jm-2 for austenite/pearlite interface, 0.7 Jm-2 for austenite/austenite interface [13], the grid spacing x, which is taken
to be 0.2 m, and the interface thickness , which in this work equals to 1 m.

Page 6 of 13

Rudnizki et al. Integrating Materials and Manufacturing Innovation 2012, 1:3

http://www.immijournal.com/content/1/1/3

3D-Modelling Due to the fact that a 3D-simulation demands high computational capacity as well as being very time consuming, a small simulation domain of 30 30 30
m3 is selected for 3D-simulations of microstructure evolution during processing of
cold-rolled dual-phase steel. Figure 4 shows this created structure consisting of about
16% of pearlite in a ferritic matrix as a 3D-volume domain. Again, in this representation ferrite is white, while pearlite as well as the grain boundaries are black.
However, the direct comparison of 3D-structures with experimental data is hindered
due to the fact that grain size distribution obtained from the volume element and from
the surface of the same structures could be very different [25-27]. For the comparison
of the simulated 3D- and 2D-structures the 3D-volume element has been be subjected
to a series of cuts in order to get sufficient cumulative area of 2D-cuts. Figure 5 shows
nine 2D-cuts of the 3D-volume presented in Figure 4 and comparison of the obtained
ferrite grain size distribution with experimental data from EBSD-analysis.
The grain size distribution of the created volume domain obtained from the XYsections reflects an elongation of the real microstructure due to cold-rolling. Moreover,
the ferrite grain size distribution of 3D-structure obtained from 2D-cuts is in fair correlation with the experimental data. The average grain size of created 3D-structure
equals to 11.5 m and diverges only by 23 m from the experimentally-obtained grain
size value 8.6 m.
As in the case of 2D-structure, the content of alloying elements is taken according to
the microprobe results and the stored energy of ferritic grains in the created 3Dstructure is set corresponding to the EBSD results. Most simulation parameters such as
behaviour of substitutional alloying elements, mobility, interfacial energy and diffusion
parameters are taken from 2D-approach. However, the nucleation parameters are
adjusted due to the fact that continuous nucleation proposed for 2D-simulations
assumes not only nucleation of new grains but also the growing up of grains from
neighbour level. This is therefore not suitable for 3D-simulations. Thus, it was established that the site saturation nucleation is much more applicable for the 3D-modelling
of recrystallisation due to the fact that according to the experimental data, recrystallisation under the investigated heating parameters takes only 2030 s. In order to reflect
the experimentally-obtained average grain diameter after recrystallisation, which is
determined by the nucleation density, it is supposed that 50 nuclei form at 660C. The
grain junctions and interfaces are addressed as probable nucleation sites. Considering
austenite formation, as in the case of 2D-simulation it is assumed that nucleation takes

Figure 4 The generated initial structure for 3D-simulation.

Page 7 of 13

Rudnizki et al. Integrating Materials and Manufacturing Innovation 2012, 1:3

http://www.immijournal.com/content/1/1/3

Page 8 of 13

Figure 5 2D-cuts through the 3D-volume element (left), and comparison of ferrite grain size
distributions of 2D initial structures and of 2D-cuts of 3D-calculation (right).

place only in pearlite. The number of nuclei is fixed to 25 so that in each pearlitic grain
several austenite nuclei could form.

Results and discussion

In the following the results of the 2D- and 3D-modelling for the microstructure evolution during intercritical annealing with the process parameters according to Figure 2
will be presented.
Figure 6 represents 2D-simulated evolutions of transformed fractions. The data repeat quite well the experimental results for the same process conditions obtained by
image analysis of corresponding micrographs. The analysis of the data shows that the
evolution of transformed phase fractions during intercritical annealing as predicted by
the 2D phase field simulation coincides with all experimental points.
The ferrite fraction predicted by the 3D-simulation during intercritical annealing with
selected process parameters also reflects the experimental results quite well, Figure 7.
The other phases are not shown in this diagram in order to keep the clearness in the
representation. Beside the good agreement with the experimental data, evolution of the
ferrite fraction from 3D-simulation reproduces well the 2D-resuts. Finally, the ferrite

Transformed fractions

1.0
0.8
Rx Ferrite
non RX Ferrite
Pearlite
Austenite
Total Ferrite
EXP Rx Ferrite
EXP Total Ferrite

0.6
0.4
0.2
0.0
0

100

200

300
Time, s

400

500

600

Figure 6 2D-simulation results for the through process kinetics of transformations compared with
experimental data from interrupted dilatometer tests.

Rudnizki et al. Integrating Materials and Manufacturing Innovation 2012, 1:3

http://www.immijournal.com/content/1/1/3

Page 9 of 13

Ferrite fraction

1.0
0.8
PFM-2D
PFM-3D
LOM

0.6
0.4
0.2
0.0
0

100

200

300
Time, s

400

500

600

Figure 7 3D-simulation results for the evolution of the ferrite fraction evolution compared with 2Dresults and experimental data.

amount simulated in 3D and 2D is inside the scatter bars obtained by metallographic

analysis.
The comparison of 2D- and 3D-simulated and experimental grain size distributions
of ferrite after intercritical annealing is presented in Figure 8 at 650C. Both simulated
grain size distributions matches the EBSD results quite well. Moreover, the average ferrite grain sizes at 650C obtained from the 2D- and 3D-simulations versus experiment
are 7.9 m, 9.8 m and 8.0 m, respectively.
Mecozzi et al. have shown that the predicted ferrite grain size distribution depends in
detail on the assumed nucleation behaviour [23]. Further, it is suggested that the apparent mobility in 2D is lower than in 3D to match a reference kinetics for a given nucleation scenario. In the present simulation 2D and 3D mobilities are the same, while in
2D continuous nucleation has been assumed and in 3D all nuclei form at the same
temperature (site saturation). The apparent agreement of 2D and 3D mobilities might
be accepted to be a compensation effect of 2D vs. 3D growth geometries and increased
nucleation spread in 2D as compared to 3D simulations.
For the more efficient evaluation of modelling results, the quality of the phase transformation simulation is verified by comparing micrographs and simulated microstructures at selected temperatures following cycle from Figure 2. Figure 9 represents the
micrographs and outputs from 2D-simulation at 710C (after recrystallization), 800C
(heating to intercritical temperature) and 650C (subsequent cooling after intercritical

Figure 8 2D- and 3D-simulation results for the ferrite grain size distribution at 650C compared
with the experimental data from EBSD analysis.

Rudnizki et al. Integrating Materials and Manufacturing Innovation 2012, 1:3

http://www.immijournal.com/content/1/1/3

Figure 9 Nital-etched micrographs of the samples (left), 2D-simulated phase where ferrite is white
and austenite gray (center), and simulated carbon distribution (right); the top row corresponds to
710C, the center row to 800C and the bottom to 650C.

annealing). The left images represent micrographs of the nital-etched samples interrupted at the corresponding temperatures and quenched to Room temperature. In
these micrographs, ferrite is bright, and martensite (which is assumed to have been austenite before quenching) as well as grain boundaries are dark. The central and right
images represent simulated phase and carbon distribution; here the same colour scale
is used as in Figure 1.
The visual comparison of the phase distributions and the grain sizes at 710C shows
that 2D-results are quite similar with real micrograph. Both, metallographic and simulation results show that the recrystallisation has been completed. At corresponding
temperature the deformed elongated ferritic grains are substituted by the round recrystallized grains. Nvertheless, there is still room for improvement. One aspect that is not
considered in the simulations is banding due to segregation effects that is quite obvious
from the micrograph shown in Figure 1.
At 800C, the calculated austenitic fraction is about 41%, which agrees with the experimentally observed 42%. The grain distributions in the simulated structure and in
the micrograph are similar. An elongated arrangement of phases can be observed in
both as a result of the stretched arrangement of pearlite in the initial structure. Due to
the fact that the ferrite to austenite phase transformation is controlled by carbon diffusion, a redistribution of carbon during the progress of phase transformation occurs. At
this temperature, the austenite contains about 0.24 wt.-% carbon.
The simulated structure at 650C with about 77% of ferrite and 23% of austenite also
corresponds well to the real dual-phase microstructure. The experimental results with

Page 10 of 13

Rudnizki et al. Integrating Materials and Manufacturing Innovation 2012, 1:3

http://www.immijournal.com/content/1/1/3

78% of ferrite and 22% of martensite confirm the simulation. Although the phase arrangement and distribution of ferritic grains in both structures are very similar, the network of austenitic/martensitic grains in the simulated structure seems to be somewhat
coarser. This could be a consequence of the fact that the final dual-phase microstructure still reflects the initial distribution of pearlite. Though the performed simulation
based on the real micrograph, very small pearlitic grains (smaller than 3 grid element)
have not been reproduced in the starting structure. The absence of these grains may
lead to the somewhat coarser martensite network in the final simulation structure.
The carbon distribution map shows the enrichment of carbon fraction in the
remaining austenite with decreasing austenite fraction. The average carbon concentration in the austenite at 650C reached a value of about 0.41%.
Figure 10 shows the phase and carbon distributions at 710C, 800C and 650C predicted by the 3D-simulation with the same colour scale as in the 2D-results in Figure 9.
As mentioned above, the size of ferritic grains from 3D-simulation in general agree 2Dresults and experimental data, Figure 10. The 3D-results concerning phase arrangement, seem to be similar to the 2D-simulation and experimental data as well. Moreover, the average carbon concentration in the austenite at 650C according to 3Dresults is about 0.46 wt.-% that also matches 2D-simulation result.

Outlook
From the information of the carbon content in austenite, the Ms-temperature at each
point of structure can be calculated. Thus, from the structure simulated by means of
Phase-Field Method, ferrite-martensite structure can be obtained by the application of
Koistinen-Marburger [28] or an alternative approach. This will enable coupling to the
already available models for the modelling of mechanical properties, such as for example RVE-FE Method for the prediction of flow behaviour [29-31].
Conclusions
2D- and 3D-simulations of microstructure evolution during processing of dual-phase
steels from cold-rolled strips have be realized by means of Phase-Field Modelling approach. It allows to describe all metallurgical phenomena occurring on a microstructural

Figure 10 3D-simulated phase distributions (top row) at 710C (left), 800C (center) and 650C
(right) where pearlite is black, ferrite white and austenite gray, carbon distributions at
corresponding temperatures (bottom).

Page 11 of 13

Rudnizki et al. Integrating Materials and Manufacturing Innovation 2012, 1:3

http://www.immijournal.com/content/1/1/3

scale during intercritical annealing, i.e. recrystallisation, austenite formation, ferrite formation as a function of chemical composition, starting microstructure and process parameters. The accurate definition of input model parameters yields both 2D- and 3Dapproaches to simulate the microstructure evolution successfully. The comparison of
2D- and 3D-simulated results with experimental data demonstrates an overall agreement of the predicted evolution of phase fraction and grain size distribution.
Moreover, evolution of carbon distribution, especially the carbon distribution in the
final structure, is an important output of Phase-Field simulations. Carbon dissolved in
martensite determines its hardness, and therefore, mechanical properties of the investigated steel. This can be used for the following prediction of the microstructure evolution. It will allow determining martensite start temperature in case of the subsequent
quenching as well as further modelling of flow behaviour of material.
The advantage of 2D-simulation compared to the 3D-approch is the achievement of
fast results and directly comparison with experimental data enabling rapid revelation of
the influencing parameters. Therefore it can be utilised for the optimisation of process
parameters to achieve the essential microstructure. In contrast, the 3D-simulations are
much more time-consuming and need high computational capacity but can be applied
easily for the following coupling with the models for the prediction of mechanical
properties.
Competing interests
'The author(s) declare that they have no competing interests.
Authors contributions
JR did the experiments, the calculations and prepared the first version of article text. UP discussed and analysed the
experimental and simulation results, contributed during text and figure preparation and corrected the article. WB
analysed and discussed the final results and conclusions and the article text. All authors read and approved the final
manuscript.
Acknowledgement
This research was carried out under project number MC5.06257 in the framework of the Research Programme of the
Materials innovation institute M2i (www.M2i.nl). The authors acknowledge the financial support of M2i as well as the
fruitful discussion with Henk Vegter, Piet Kock (both now at Tata Steel Europe) and Jilt Sietsma.(Delft University of
Technology).
Received: 4 July 2012 Accepted: 6 August 2012
Published: 31 August 2012
References
1. Bag A, Ray KK, Dwarakadasa ES (1999) Influence of martensite content and morfology on tensile and impact
properties of high-martensite dual-phase steels. Metallurgical and Materials Transactions 30A 5:119311202
2. Tomita Y (1990) Effect of morphology of second-phase martensite on tensile properties of Fe-0.1C Dual-phase
steels. Journal of Materials Science 25:51795184
3. Militzer M, Poole WJ (2004) A critical comparison of microstructure evolution in Hot-rolled and cold-rolled dualphase steels. AHSSS Proceedings :219229
4. Huang J (2004) Microstructure evolution during processing of dual phase and TRIP steels. PhD Thesis, University of
British Columbia
5. Peranio et al (2010) Microstructure and texture evolution in dual-phase steels: competition between recovery,
recrystallization, and phase transformation. Materials Science and Engineering A 527:41614168
6. Calcagnotto et al (2008) Ultrafine grained ferrite/martensite dual phase steel fabricated by large strain warm
deformation and subsequent intercritical annealing. ISIJ International 48:1096
7. Calcagnotto et al (2011) and fracture mechanisms in fine- and ultrafine-grained ferrite/martensite dual-phase
steels and the effect of aging. Acta Materialia 59:658670
8. Bker M (2002) Numerische methoden in der materialwissenschaft. Braunschweiger Schriften des Maschinenbaus
8, Braunschweig ISBN 3-936148-08-2
9. Raabe D, Roters F, Barlat F, Chen LQ (2003) Continuum scale simulation of engineering materials. Weinheim WileyVCH Verlag, Weinheim
10. National Research Council (2008) Integrated computational materials engineering: a transformational discipline for
improved competitiveness and national security. National Academic Press, Washington D. C
11. Gottstein G (2007) Integral materials modeling. Weinheim Wiley-VCH-Verlag, Weinheim

Page 12 of 13

Rudnizki et al. Integrating Materials and Manufacturing Innovation 2012, 1:3

http://www.immijournal.com/content/1/1/3

12. Rudnizki J, Bttger B, Prahl U, Bleck W (2011) Phase-field modelling of austenite formation from a ferrite plus
pearlite microstructure during annealing of cold-rolled dual-phase steel. Metallurgical and Materials Transactions A
8:25162525
13. Thiessen RG (2006) Physically-based modelling of material responce to welding. PhD Thesis, TU Delft
14. Mecozzi MG (2007) Phase-field modelling of the austenite to ferrite transformation in steels. Ph.D Thesis, TU Delft
15. Militzer M (2011) Phase field modeling of microstructure evolution in steels. Current Opinion in Solid State and
Materials Science 15(3):106115
16. Zaefferer S, Konijnenberg P, Demir E, Woodcock T (2010) Progress in 3-dimensional EBSD-based orientation
microscopy: New software tools for 3-dimensional materials characterization. Materials Science and Engineering
MSE 2010, Darmstadt
17. Eshelby JD, Read WT, Shockley W (1953) Anisotropic elasticity with applications to dislocation theory. Acta
Metallurgica 1:251259
18. Steinbach I, Pezzolla F, Nestler B, Seeelberg M, Prieler R, Schmitz GJ, Rezende JLL (1996) A phase field concept
for multiphase systems. Physica D 94:135147
19. Eiken J, Bttger B, Steinbach I (2006) MultiPhase-field approach for multicomponent alloys with extrapolation
scheme for numerical application. Phys Rev E 2006:066122
20. Steinbach I (2009) Phase-field models in materials science; a tutorial review. Modelling and Simulation in Materials
Science and Engineering 17:07300131
21. MICRESS The Microstructure Evolution Simulation Software http://micress.de
22. Thermo-Calc Software http://www.thermocalc.com
23. Mecozzi MG, Militzer M, Sietsma J, van der Zwaag S (2008) The role of nucleation behavior in phase-field
simulations of the austenite to ferrite transformation. Metall Mater Trans 5:12371247, A 39
24. Krielaart GP, van der Zwaag S (1998) Simulations of pro-eutectoid Ferrite Formation using a Mixed Control Growth
Model. Material Science and Engineering A246 1998:104116
25. Giumelli AK, Militzer M, Hawbolt EB (1999) Analysis of the austenite grain size distribution in plain carbon steels.
ISIJ International 39:271280
26. Calcagnotto M et al (2010) Orientation gradients and geometrically necessary dislocations in ultrafine grained
dual-phase steels studied by 2D and 3D EBSD. Mater Sc Engin A 527:2738
27. Zaefferer S et al (2008) Three-dimensional orientation microscopy in a focused ion beam-scanning electron
microscope: A new dimension of microstructure characterization. Metal Mater Trans A 39A:374389
28. Koistinen DP, Marburger RE (1959) A general equation prescribing the extent of the austenite-martensite
transformation in pure iron-carbon alloys and plain carbon steels. Acta Metallurgia 7:5960
29. Rodriguez R, Gutierrez I (2004) Mechanical behaviour of steels with mixed microstructure. Proceeding of TMP04,
B-Liege 363:356363
30. Thomser C, Uthaisangsuk V, Bleck W (2009) Influence of martensite distribution on mechanical properties of dual
phase steels: experiments and simulation. Steel research international 80(8):582587
31. Uthaisangsuk V, Prahl U, Bleck W (2009) Failure modeling of multiphase steels using representative volume
elements based on real microstructures. Procedia Engineering 1(1):171176
doi:10.1186/2193-9772-1-3
Cite this article as: Rudnizki et al.: Phase-field modelling of microstructure evolution during processing of coldrolled dual phase steels. Integrating Materials and Manufacturing Innovation 2012 1:3.

Submit your manuscript to a

journal and benet from:
7 Convenient online submission
7 Rigorous peer review
7 Immediate publication on acceptance
7 Open access: articles freely available online
7 High visibility within the eld
7 Retaining the copyright to your article

Submit your next manuscript at 7 springeropen.com

Page 13 of 13